Volume 23, Issue 4, Pages 762-773 (April 2015) Probing the Protein Interaction Network of Pseudomonas aeruginosa Cells by Chemical Cross-Linking Mass Spectrometry  Arti T. Navare, Juan D. Chavez, Chunxiang Zheng, Chad R. Weisbrod, Jimmy K. Eng, Richard Siehnel, Pradeep K. Singh, Colin Manoil, James E. Bruce  Structure  Volume 23, Issue 4, Pages 762-773 (April 2015) DOI: 10.1016/j.str.2015.01.022 Copyright © 2015 Elsevier Ltd Terms and Conditions

Structure 2015 23, 762-773DOI: (10.1016/j.str.2015.01.022) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 1 In Vivo Cross-Linking-Derived Protein-Protein Interaction Network of P. aeruginosa PIR cross-linked derived large-scale PPI profile obtained from intact PA cells (A) and a peptide-centric view of the major lipoprotein interactions (B). Nodes in (A) represent proteins and connections denote cross-link-interactions. In the peptide-centric view of the lipoprotein interactions (B), multiple (≥2) cross-linked peptides of a single protein are arranged in a circle layout. Edge colors in (A) and (B) represent computationally predicted PA protein interactions from PAO1 interactome database (red), homologous PPI in E. coli as reported in the EcID database (blue), homologous PPI from cross-linked E. coli cells (green), and novel interactions (black). Intra-links in (A) are shown as semicircular edges and the homodimeric links in (B) are indicated by colored semicircular (magenta) edges. Node colors from (A) and (B) denote protein subcellular locations. The PPI network layout was directed by connectivity of the nodes in the Cytoscape program and the three highly connected membrane proteins, namely OprI, OprL, and OprF, are highlighted by bigger node size. A pie chart from (A) shows the distribution of the subcellular locations of all the 211 identified cross-linked proteins. Structure 2015 23, 762-773DOI: (10.1016/j.str.2015.01.022) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 2 The Identified Inter-Cross-Linked Sites of the Major Lipoproteins OprI, OprF, and OprL A block diagram shows relative protein lengths of OprI, OprF, and OprL with black lines denoting the positions of all Lys residues within the mature form of lipoprotein sequences. Dotted and solid black lines mark the positions of un-cross-linked and cross-linked Lys. Homodimeric sites are marked in red. Pink, orange, and light blue lines denote inter-cross-linking between OprI-OprF, OprI-OprL, and OprF-OprL, respectively. The OmpA-like region in OprL and OprF is shown as a shaded blue box. See Figure S2. Structure 2015 23, 762-773DOI: (10.1016/j.str.2015.01.022) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 3 Known and Cross-Linking-Derived Predicted Structures of Fused Form of SecD and SecF Proteins from Bacteria Thermus thermophilius and P. aeruginosa, Respectively (A) Crystal structure of the fused SecD-SecF protein (TsecDF) from Thermus thermophilius showing the SecD and SecF domains containing six transmembrane domains (yellow) each, and periplasmic (gray) and cytoplasmic (purple) domains. Location of the cross-linked peptides from PAO1 SecD (red) and SecF (blue) was mapped onto the crystal structure. (B) The top scoring predicted structure of concatenated PA SecD-secF using I-TASSER. The green and gray portions denote the SecD and SecF portions of the concatenated complex. See Figure S3. Structure 2015 23, 762-773DOI: (10.1016/j.str.2015.01.022) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 4 Prediction of OprI Monomer Structure Using the Template Structure of Homolog Protein LPP and Cross-Linking-Aided Selection of OprI Homodimer Complex Model (A) Sequence alignment of OprI and its E. coli homolog LPP showed 25% sequence alignment. Conserved amino acid residues (gray), lipid-bound cysteines (cyan), the signal peptide sequence at the N-terminal (purple) are highlighted. Lys79 from OprI found inter- and intra-cross-linked is highlighted in red. (B) OprI monomer structure was predicted using I-TASSER, using LPP-56 mutant protein as a template. The cross-linked Lys molecules are shown on the ribbon structure of OprI. (C and D) Two monomers of OprI were docked onto each other using SymmDock to generate a dimer, with (C) and without (D) using the cross-linking site distance constraint information. The resulting dimer models were evaluated for the distance between the cross-linked Lys. See Figure S4A and S4B. Structure 2015 23, 762-773DOI: (10.1016/j.str.2015.01.022) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 5 Sequence Alignment of OmpA-Like Domains of OprF and E. coli Homolog OmpA, Showing Locations of Cross-Linked Sites and the Top Selected OprF Homodimeric Model, Based on the Cross-Link-Derived Distance Constraints (A) Sequence alignment of the OmpA-like regions of PA OprF (black) and E. coli homolog OmpA (red). The sites of homodimeric intra-links in OmpA identified by in vivo cross-linking (Zheng et al., 2011; Weisbrod et al., 2013a) are indicated by dotted arrows, and the sequence region reported to be important for OmpA dimerization (Marcoux et al., 2014) is underlined (dotted line). All cross-linked Lys residues of OprF are shown in bold. The peptide sequences containing sites of homodimeric intra-links (K248, K250, and K291) are underlined by solid lines and the corresponding cross-linked Lys residues are marked by solid arrows. Putative PG-binding sites on OprF (residues 284–297) are shown in blue. (B) Top OprF dimer model with the shortest Xwalk distance between K248 (red), K250 (cyan), and K291 (yellow). The space filled region highlights the predicted dimer interface and the aligned PG-binding site. (C) Xwalk distances between Lys residues are shown in the table under the rotated view of the dimer model. See Figures S4C and S4D. Structure 2015 23, 762-773DOI: (10.1016/j.str.2015.01.022) Copyright © 2015 Elsevier Ltd Terms and Conditions